AI Leveling the Municipal Playing Field

In this edition of MMA’s Potholes & Politics hosts Rebecca Lambert and Amanda Campbell are joined by Audra Caler, Camden Town Manager; Kyle Hadyniak, Gray IT & Communications Director; Brian McDonald, MMA IT Director, and Anthony Wilson, Winthrop Town Manager to discuss how AI can support municipal work and services.  The key takeaway from this session is that with practice, policies and responsible use, AI can level the playing field among municipalities by providing access to cost effective professional resources.  

Transcript

Welcome everyone to Potholes and Politics, Local Maine Issues from A to Z. I'm your co-host Rebecca Lambert, and with me as always is my amazing colleague, Amanda Campbell. 

Good morning, everybody. Thanks, listeners for joining us. 

Yes, thank you. In the last episode, we had a great conversation with Representative Liz Caruso from Caratunk and Representative Mike Ray from Lincolnville, and they shared their insights about serving at the local level compared to their first session as state legislators. They told us all about their successes and challenges, and we would encourage our listeners to go back and listen to that episode if you haven't already. 

And while you're at it, don't forget to like and subscribe to our podcast on your favorite platform. By doing that, you'll be notified when a new episode drops, ensuring you'll never miss one of these compelling and always entertaining episodes.

We are nothing if not entertaining. 

We are entertaining.

This week we are welcoming MMA's Director of IT and Administration, Brian McDonald, who will be chatting about artificial intelligence and how municipalities can think about integrating AI to streamline workflows and create efficiencies.

We're also welcoming Anthony Wilson, Town Manager of Winthrop, Kyle Hadyniak, Director of Communications and IT for the Town of Gray and Audra Caler, Town Manager of Camden. Welcome everybody and thank you so much for joining us. 

Thank you. 

Great to be here. 

So, let's start this conversation off with you, Brian. Tell me about AI policy. Why is it important to have one? 

So, what we, I think determined was that a lot of people were jumping head first into AI, and, maybe there were some items that weren't understood or there were some potential pitfalls that we wanted to make sure people understood. So, we saw it as an opportunity, for really, for education, but also for protection from specific liabilities.

So the use of AI presents a lot of opportunities, but also challenges, potentially with protected data, potentially with malicious use. So, we wanted to be sure that members  had the understanding of what could go wrong so that they could implement successfully and realize not only those kind of gains and efficiencies but also be protected when doing so.

That makes a lot of sense. And Anthony, we knew that you helped to develop Winthrop's AI policy. How did that get started and how has its implementation gone? 

Well, it got started because I attended a conference that Brian was a keynote speaker at, and it was, developing a policy was one of two main takeaways that I had from that conference.

And so, I contacted Brian afterwards. There is a template on the MMA website that we worked off of, and we just began going through it and trying to determine sort of, what best fit our needs. I presented something to our council a few weeks ago and they decided to sort of pump the brakes on that, so we haven't actually implemented it.

And the reason for that is there is a statewide task force that is addressing AI. And it's gonna be issuing some recommendations in October. And so, our council decided they wanted to see what those recommendations entailed before they approved our AI policy just in case there was something from those recommendations we might want to incorporate.

And we're not sure that there's actually gonna be anything of value for municipalities within those recommendations because the charge of that group is very expansive. But I completely agree with Brian. We need to make sure there are some safeguards to just ensure that we are using AI in a responsible manner that not only improves our efficiency, but protects the information, particularly sensitive information that we might be using as part of that process.

In what areas do you see AI being most useful to you in your municipal work? 

Well, so my background is, I was a journalist for 20 years. So, I'm a very good writer. But what I have found is that by using AI to review important memos and sensitive emails and grant applications and such, it has helped to refine the clarity, condense the language somewhat and consolidate some points in ways.

I think that just make that a lot more accessible to people.  I always write my own drafts and then feed them through AI to have that refined so that I'm not de-skilling myself by just saying, “Hey, ChatGPT, will you do this for me?” It's always, I create the content and then ask AI to look at refining it.

And one of the really valuable things that I've used AI for thus far is, we were hiring for a position. We got a ton of resumes, I mean, probably three dozen resumes. I went through all those resumes myself and screened the ones that I thought were worthy of an interview. But nonetheless, after that, I fed all those resumes into AI along with the job description and it ranked the applicants for me.

And so, its ranking was very, very similar to my rank. And so, it just reinforced that I wasn't overlooking someone who had a skillset that would've been good for that job. And, at the end of the day, we ended up hiring the person who was not only at the top of my list, but at the top of the ChatGPT’s list.

So, I thought it was just a really good way to ensure that, you know, I wasn't, as I said, overlooking anyone that we really ought to be giving serious consideration to. 

That's really interesting. 

It's very cool. 

Brian, Anthony mentioned a policy template. Is that something that's available to all of our members through the website?

Yep. So, it's on the cyber portal so memun.org/cyber. It's actually part of our AI Toolkit. So, if you go to the toolkits page and it includes both the presentation that was given, which includes all of, like the risk points and risk mitigation ideas, and also the policy guidelines with the actual sample policy guidelines, on the web for any member who logs into the site.

So Audra, how about in Camden? Are you using AI in Camden and how is that going for you? 

Yeah, there are, there are just so many different use cases, especially across an organization like a local government where our business centers are so diverse. So, I guess starting at the top assessing is looking at AI tools to improve how they conduct their work so that there's less of a need for some of the more comprehensive, list and measure revals and, just keeping that data updated on an annual basis has been really important and helpful, for assessing work. 

Our public works department uses a street scan system, which incorporates AI in making analysis about condition of the surface of the streets as well as integrating things like, the life of the subsurface infrastructure so we can make more informed decisions about big capital investments in repairing the entire roadway all the way from subsurface up to resurfacing. I use it all the time for a number of different applications.

So, Anthony was talking about it's great for HR purposes, it can be very helpful with grant applications. I think all of us are in a position where we get these, large technical reports from consultants and trying to distill that information in a more digestible manner for elected officials or for members of the public.

I've found AI to be very helpful in condensing it into, like a succinct memo that, of course people have access to the full report. But having that really, more short, condensed version is helpful to get people across what they need to make decisions about or what they need to talk about.

And, you know, it's finance. I see a lot of potential there as well in analyzing data over number of years to look at trends. So. Really the use cases are, there's no limit to what it can do. It's really the limit of what we can imagine it doing. And, I found it was really helpful. I just got back from the three week Harvard Kennedy School state and local government executive education course and there was a big focus on AI.

And at the Harvard Kennedy School, they're very concerned that the public sector is going to be left behind. And that's a concern that I have particularly in local government because, one of the assertions that they're making, and I completely agree with it, it's going to be like somebody who can do a, a good Google search 15 years ago, 20 years ago.

So it's, it's gonna be that ubiquitous that everybody's going to have to be comfortable with these tools and proficient with them. And so, I think that, practicing with it and getting better at using it and getting more creative about how you can use it, especially people like us who are further along in their careers.

So we have the ability to verify that what it's producing is accurate or inaccurate is gonna be really important for everybody in our industry. 

Absolutely. Someone just told me recently that AI won't take your job, but someone who knows how to use AI will take your job. 

I, and I think that's true. Like any, any technology that comes out, it's gonna be really important supplement to how we do our jobs.

Exactly. So Kyle, as a fellow IT Director, how have you been supporting municipal officials in Gray with the implementation of AI? 

Yeah, so I'm also of the mind, I'm also of the concern that municipalities may get left behind when it comes to the AI revolution. And so, what I'm trying really hard to do is make sure employees are at least aware of what generative AI can do for them.

And you know, you can't really force them to adopt a new technology 'cause that's like the way to get them not to want to use it. Right? So, one of my jobs that I'm tackling every day is selling employees on this new technology. And luckily in Gray, I'm very happy with our workforce. It generally tech forward.

Our town manager is tech forward. He supports a lot of these tech-focused initiatives that I bring to him and to other members of staff. So, getting AI in the conversation here in Gray was not hard and some of employees are really excited about AI like that. So mainly how we have adopted generative AI tools is through the Microsoft 365 Suite.

Town of Gray is firmly a Microsoft shop as are, I'm sure, most municipalities. And if you live in that environment, you know Microsoft Copilot is like the generative AI tool to use because it can be built into those 365 applications. This is really important because then employees who use those applications every day already have a familiar interface with which to use generative AI.

And we've been using copilot specifically in a variety of ways. And I'll mention first the main way that I love using the generative AI capabilities of Copilot, and that's in Microsoft Teams. So here in Gray, we use Teams for all of our internal and external meetings. We don't use Zoom, because we already pay for Teams.

So, you know why pay for both? Copilot being built into teams is supremely useful. It will provide you AI summaries of meetings, including like town council meetings, which could be two or three hours long, and to have an AI generated recap of that meeting available within minutes after it ends is, it has allowed me to create public facing recap articles that I publish on our website.

And of course, that's not to say that I trust Co-pilot's output at face value. Having a human in the loop is still incredibly important. You should never take generative AI's output at face value. Always check it yourself. But it saves me a lot of time. It saves me hours from having to go through a three-hour meeting myself and pull out the salient bits and put it into an article.

So that is one major way that I, I love using AI and generally, I'm enormously optimistic about AI. A lot of people may have some notable concerns that will be addressed in time. But otherwise, I am completely on board and I'm trying to get others on board. You could, you could blame it on my lifelong love of Star Trek that I love generative AI and want to see it succeed, especially in a municipal workplace.

So, if each of you were to say one thing to officials in the community that might be on the fence about introducing AI, what would your recommendation to those folks be? 

So, I'll go first. The, the main piece of advice I would give is you can't afford to miss out. It's like, not bothering how to use a web browser or learning how to use a computer, back in the day.

I completely agree with the sentiment shared earlier that a human, that AI will not replace you, but it will, someone using AI will replace you. I've heard that for a while now, and it's just increasingly truer and truer. You can't afford to miss out, it's probably easier than you think to learn about generative AI tools.

I mean, go to YouTube University and start learning yourself. Microsoft has a really handy website called Microsoft Learn, where you can learn all about their products, including Copilot. But ultimately you can just ask a generative AI tool like ChatGPT, or Gemini or Copilot yourself. Just ask it. How can you help me? I'm a municipal official in this town. Here's what my work entails. How can you help me? 

And because it's able to understand natural language and then output natural language, it could tell you specifically like, “Hey, I could help you, create emails, or refine language, or create posters, or do meaning recaps”, all that good stuff. It'll tell you yourself. So, just don't miss out. Thank you.

I mean, I would say the same. I, I think that there's a lot of concern about what could happen and a lot of, I, I've seen, a lot of people, like, we, we led with, you gotta have a policy. And I'm not saying that you don't have to have a policy. I'm just saying I think that people are gonna use it regardless of whether you have a policy or not.

So, I think that the, the biggest thing that you can do is, especially with municipal staff, is encourage them to practice with it, get good at learning how to prompt it, experiment with it. Because if you don't have proficiency with it, it's gonna be a bigger problem in the end than waiting until you have the perfect policy and all of your elected officials are on board.

Because they often don't really understand the implications of it for your work, necessarily. So, I just think it's gonna be so important that people understand how to use it. And by understanding how to use it, you can also explain better to elected officials that might be making decisions about it or community members that have concerns with it, what the actual use cases are in a municipality and what it's gonna mean for everybody. So, I think that the caution, completely understand it. You know, we all understand policy, we understand operating under policy. However, I don't think letting that be sort of a barrier to experimenting and practice and just figuring out all the different ways that it could augment what you're doing.

Yeah, Kyle sort of stole my thunder because, at the conference where I mentioned that, that I first heard Brian speak about, AI policies, there was a fellow who served on a very small town select board there who said that there were people in his community who were saying that their town is just too small to be using AI.

And I just sort of chuckled to myself thinking that's like saying, oh, you know, this internet thing and computers, you know, we're too small to be tapping into, into that sort of thing. It is coming, in fact it's already here. And so, it is very important to, to get on board and to Audra's point, earlier, I read this really interesting article in the New York Times over the weekend where, CEOs are pushing AI, but they don't really understand it or using it themselves. And so now there's this push to ensure that leaders are not just doing, as I say, but doing as I do, and sort of leading by example. And so, I think it's really important for those of us who are in municipal government leadership positions, to model, sort of the behavior that we want people to adopt in order to operate more efficiently.

And I have not once sat down with one of my colleagues and showed them how I was using ChatGPT, that they weren't just like amazed and excited about improving an email or a background memo. And it literally takes, like, four seconds for ChatGPT to spit out a better version of what you've created.

But in the same regard, sometimes I do that and I think, you know what? I like my original wording better. And, and again, to Kyle's point, you've got to review whatever it spits out and just don't take it and cut and paste it into whatever work product that you're working on. It still takes a human brain in order to figure out what is best given the circumstances of whatever it is you're working on.

Anthony on that, on your point real fast here. I also heard the other side of the argument saying, Gray might be too small of a municipality to worry about this stuff. But I've always considered generative AI and the concerns surrounding it, a lot like the safe sex talk you might give teenagers where you know they're going to do it, but you just wanna make sure they do it safely.

So that's the idea of adopting a policy and getting it out to the staff and to, even the awareness of the community at large. Like they would probably be, they would probably feel great if they knew their municipal staff were being safe with this new technology. 

Yeah. Brian, at that conference helped us sort of demystify AI to some extent by saying it is all around us and we are using it all the time. Whether you actually, consciously know that or not. Google is AI. Your Netflix list is AI. Every time you go onto Amazon, it's AI. So, we're constantly surrounded by it. It's just, it's a little bit more subtle, I guess, than, than using a specific AI platform.

Any parting thoughts, Brian? 

Yeah, so, this is great examples and great comments. So that I have also heard from several members that we are too small. We have a staff of two people. We have a population of 400 people. There's a really kind of unique advantage for AI in Maine because everywhere in Maine, every municipality has to provide the same, I guess, services, right? To their communities. And so, when you're talking about a community of 500 people with maybe two municipal staff, you're still covering the same amount of roads and area and emergency services. And so, in, in the past, those communities have not had potentially, let's say, the funding to bring in a consultant to do that street kind of work or maybe, sewer assessments or assessing analytics and AI tools make that really possible now for really, really good price. And so, when you start looking at things like the assessing or the street reviews or the smart city management, smart city planning, I know there's several municipalities now using it for like water quality assessment and sewer assessment.

Talk about a force multiplier for a relatively reduced rate, where the days of having to go and get a consultant for a couple hundred thousand dollars, AI can potentially analyze and give you a pretty good plan in seconds. So that's what I would say. There is no too small, and in fact, I would argue that there may even be an additional benefit for smaller communities that larger communities maybe wouldn't recognize because they have individuals on staff who might not want to, go that route or they have, people kind of doing that assessing all the time. So, getting everybody from large to small to utilize those AI tools. Everybody can realize the benefit, but the small community, I think there's a really untapped benefit there that needs to be explored.

Hmm. 

I, I agree. I think that's where the biggest potential is because they have the least resources. So, this really is gonna optimize their capacity at a much greater rate than the more well-resourced places. So, it's a I, I just think it's a fantastic opportunity because the majority of municipalities in Maine are very, very small and under-resourced, and they're expected to do just so much with so few people that have the training to do these things. So, this could be really game changing for them in tackling some of their more complicated problems. 

I always tell employees generative AI should be, you should treat it like an assistant. And if you have an assistant on staff that costs little to no money and doesn't require benefits or time off and it can operate 24 / 7, then just think of how that could help a municipality of like five employees be able to tackle all the things they need to tackle.

It's common for municipal workers to wear many hats and generative AI can help fulfill quite a few of those responsibilities. And in, indeed, it is really just the imagination is the limit. I'm constantly amazed, even though I work with AI every day still what it's capable of, what reasoning it can do, and what content it can generate, and how it can make your work life easier.

So, for those, five, person municipalities, a 10-person municipality, uh, you know, workers, uh, don't, don't miss out, please. 

It'll just make your life so much easier. 

Yeah. 

Yeah. 

So, clearly our guests are well-versed in AI. Brian, what sorts of tools does our AI Toolkit on the member website offer to those members who may not be as well-versed as our guests?

So, we, we start the same way I think every AI conversation started. We start really with governance. So, we are pushing the governance education tools. We entered into a collaborative partnership with InnovateUs to provide free, AI training to local governments, local entities. That training is specific to local government and it's being constantly developed with additional workshops, I think, they're dropping 'em on a weekly or a biweekly basis.

So that brings in the educational component that I think every municipal employee, not just decision makers, should look at and it's free. But then we bring about the policy governance and kind of the risk mitigation pieces. So, we've started that as phase one. Phase two is definitely to get into more of the use cases and what municipalities are using.

We, we've talked about going out into communities and certainly, the panel we've kind of put together here is a really good representative panel of what is possible. So, I think that we're looking more for, to get into some more of that collaboration to expose more of those use cases to other municipalities and have that kind of municipal to municipal relationship foster that piece. So, we will, we'll steer heavily at MMA on the governance and education and work with some of our members to spread the phase two use cases, specifically outbound.

Great discussion. 

Yes. This has been a very interesting discussion. Thank you all for joining us today. 

Of course. 

My pleasure. 

Thank you. 

For our, listeners, we're gonna drop the links to a couple of these resources that we've talked about in our show notes, so make sure that you check that out.

I can offer a testimony about that InnovateUs program, 'cause I, I did the generative AI one and it is fantastic. It's really accessible, really educational, and it doesn't take too long to go through it and you'll learn a lot. So, I highly recommend. 

Fantastic. 

And it's, and it's free and it's great. 

You can't beat free. 

Well, and what I'll say too on that is, stay tuned because they recently have just dropped a couple of additional trainings. One specifically with AI and police use cases, and I know like there are some really targeted areas that are starting to become emergent, you know, used with police, used with assessing and, so I, I don't think we can neglect the education there as well.

So even if you have somebody who traditionally isn't using technology or, as maybe a law enforcement or a first responder, like there are like tools and components that this may also expose them to that would be useful to look at. So, the InnovateUS training isn't just a one and done, it's something that municipal leaders should really be reviewing and determining who on their staff may benefit from specific workshops and trainings.

And I would check back every week. It doesn't look like it's gonna slow down. But just look for areas that traditionally might have been left out of the conversation. There may be some really good ways to maximize in areas that you wouldn't have traditionally thought AI could help with.

There we are then 

There we are then.

That wraps up another episode of Potholes & Politics. Thank you guys so much for joining us and for sharing your AI expertise. We appreciate you. 

Thank you. 

Thank you.

For more information about MMA’s AI Toolkit, visit www.memun.org/cyber (this will prompt you to login with your MMA member credentials.)